Cross-Linked Bacterial Cellulose Networks Using

Glyoxalization

Franck Quero1, 2, Masaya Nogi3, Koon-Yang Lee4, Geert Vanden Poel5, Alexander Bismarck4,

Athanasios Mantalaris4, Hiroyuki Yano3 and Stephen J. Eichhorn1, 2*

1Materials Science Centre, School of Materials, University of Manchester, Grosvenor Street,

Manchester, M14 9PL, UK.

2 The Northwest Composite Centre, University of Manchester, Paper Science Building, Sackville Street,

Manchester, M14 9PL, UK.

3 Research Institute for the Sustainable Humanosphere, Kyoto University, Uji, Kyoto 611-011, Japan.

4 Polymer and Composites Engineering (PaCE) Group, Department of Chemical Engineering, Imperial

College London, South Kensington Campus, London, SW7 2AZ, UK.

5 DSM Resolve, Thermal Analysis Group, P.O. Box 18, 6160 MD Geleen, The Netherlands.

*s.j.eichhorn@manchester.ac.uk

Telephone: +44 (0)161 306 5982, Fax: +44 (0)161 306 3586

ABSTRACT

In this study we demonstrate that bacterial cellulose (BC) networks can be crosslinked via

glyoxalization. Modified BC networks are found to not dissolve in a typical cellulose solvent

comprising a DMAc/LiCl mixture, whereas unmodified BC networks readily dissolve. The crystal

structure of unmodified BC networks was investigated and found to not significantly change after

glyoxalization, indicating heterogeneous modification occurs. No significant difference in moisture

content between unmodified and glyoxalized BC networks is found using thermogravimetric analysis.

However the onset of degradation and the peak temperature is found to be significantly reduced after

glyoxalization. Contact angle and zeta potential measurements confirm the hydrophobic nature of

glyoxalized BC networks. Water absorption capacity shows that unmodified BC networks have a higher

absorption capacity than glyoxalized BC networks, thus revealing their higher moisture resistance.

Fracture surfaces of samples shows that, in the dry state, less delamination occurs for glyoxalized BC

networks compared to unmodified BC networks, suggesting that covalent bond coupling between BC

layers occurs during the glyoxylation process. In the wet state, glyoxalized BC networks experience

significant delamination, but the layered structure is preserved. Young’s moduli of dry unmodified BC

networks do not change significantly after glyoxalization. The stress and strain at failure are however

dramatically reduced after glyoxylation. The wet mechanical properties of the BC networks are also

improved by glyoxylation. Raman spectroscopy is used to demonstrate that the stress-transfer of

deformed dry and wet glyoxalized BC networks is significantly increased compared to unmodified

material. This enhanced stress-transfer within the networks is shown to be as a consequence of the

covalent coupling induced during glyoxylisation, and offers a facile route for enhancing the mechanical

properties of BC networks for a variety of applications.

KEYWORDS Raman spectroscopy, bacterial cellulose, glyoxal, cross-linking, stress-transfer.

Introduction

Cellulose is the most abundant and utilized polymer on earth. It is found in higher plants, algae and is

produced, under culturing conditions by bacteria, and by sea animals called tunicates1, 2. Bacterial

cellulose (BC) is an extra-cellular product of bacteria identified for the first time by Brown3, 4. Several

families of bacteria are cellulose producers, namely, Acetobacter, Entrobacter, Rhizobium,

Agrobacterium, and Sarcina5. The acetic acid bacteria Acetobacter xylinum (reclassified as

Gluconacetobacter xylinum) have been found to be one of the most productive6 and consequently are

the most studied species. This form of BC is produced in a very pure form. Contrary to cellulose from

plants, BC is free of wax, hemicellulose and lignin7.

Cellulose has been found to have desirable mechanical properties and is consequently a good

candidate as a reinforcement in composite materials8-13. In plants, it naturally plays a structural role

owing to its high stiffness. The crystal modulus of cellulose has been found to be ~138 GPa using X-ray

diffraction14, 15. More recently Young’s modulus of single tunicate whiskers and BC nanofibers have

been estimated using Raman spectroscopy; values of 143 GPa2 and 114 GPa16 were found respectively.

Another estimation, using an AFM cantilever method, yielded a value of 78 ± 17 GPa for BC

nanofibrils17.

BC fibrous networks produced using different culturing times have been used to reinforce poly(L-

lactic acid) (PLLA)18. Observations of the fracture surfaces of samples deformed in tension indicated

that delamination occurred mainly between layers within the BC networks, rather than at the interface

between the BC and PLLA. It was also found that the more layered the structure of these networks, the

lower the stress-transfer efficiency. These layers in BC networks have been recently reported to be

weakly linked19. Consequently in order to improve the stress-transfer efficiency of BC networks, we

have crosslinked the material using glyoxal. We report the influence of this facile treatment on the

mechanical and stress-transfer properties in both the dry and wet states.

It is possible to functionalize cellulose owing to the presence of a large number of accessible

hydroxyl groups along its backbone structure. These hydroxyl groups can be functionalized via

esterification13, polycondensation, etherification or acetalyzation reactions20, to name a few. By using an

appropriate chemical substance having at least two reactive groups that can react with hydroxyl groups,

covalent bonds can be formed between cellulose polymer chains, and therefore potentially between

weakly linked BC layers. Several chemical substances have been reported as cross-linking agents for

cellulose; namely glyoxal and its homologues, but also formaldehyde, epichloroydrine, epoxides

(1,2,3,4-diepoxibutane), dichloroethane or diisocyanate20. Most of these chemicals are not

environmentally friendly, and/or toxic. Glyoxal is however an almost fully biodegradable and low

toxicity chemical (compared to formaldehyde), and can be obtained from renewable resources21 by the

oxidation of lipids and as a by-product of biological processes22. Consequently this molecule is a good

candidate to be used as an environmentally friendly cross-linker for cellulose. The main application of

glyoxal thus-far has been in the textile industry to impart durable wet mechanical properties of cotton

fabrics23. No application of glyoxal has, to the authors’ knowledge, been proposed for cross-linking BC

networks. When cellulose is exposed to water, its intra and intermolecular hydrogen bonding is

disturbed leading to a dramatic decrease of the mechanical and stress-transfer properties. By cross-

linking cellulose polymer chains with covalent bonds through glyoxalization, the effect of this

disruption of hydrogen bonding on physical properties can be reduced. We show, for the first time, the

potential of glyoxal to enhance the stress-transfer efficiency of BC networks, and thereby their

mechanical properties. This modification is shown to persist in the wet state, and so could be used for a

variety of applications; for example in composites manufacturing and biomedical applications.

Experimental Methods

Materials and chemicals Gluconacetobacter xylinum (no. 13693; National Institute of Technology

and Evaluation, Tokyo, Japan) and Hestrin-Schramm (HS)24 medium were used to produce BC

networks. The cells for the inoculum were cultured in test tubes statically at 27 ºC for 2 weeks. The

thick gel produced during culturing was then squeezed aseptically to remove the embedded cells. The

cell suspension (25 ml) was then transferred as an inoculum for the main culture (500 ml of medium),

which was incubated statically at 27 °C for 14 days. BC networks (35 mm in diameter) were purified by

boiling with 2% NaOH for 2 h, and then by washing with distilled water, followed by hot pressing at 2

MPa and 120 ºC for 4 min to completely remove the bulk water.

Glyoxal (~40 wt. % in de-ionised water), aluminium sulfate hexadecahydrate (Al2 (SO4)3 16 H20,

purity ≥ 98%), lithium chloride (LiCl; ≥ 99.5 %) and N,N-dimethylacetamide (DMAc; ≥ 98 %) were

purchased from Sigma-Aldrich.

Glyoxal treatment of BC networks A solution of 5 wt. % of glyoxal was prepared by diluting a

commercial glyoxal solution (40 wt. %) with de-ionised water. Aluminum sulfate was added with a

concentration of 1 g.L-1 to catalyze the reaction. The solution was agitated under mechanical stirring

until complete dissolution of the catalyst. Narrow strips of BC (~ 30×1×0.060 mm) were immersed in

the glyoxal solution for 5 hours. During that time the glyoxal molecules adsorbed on the surface and

impregnated the structure of BC networks. These impregnated strips were then removed from the

solution and placed in a convection oven at 150 °C for 15 minutes for glyoxalization to complete.

Modified BC networks were then washed in de-ionised water at ~70 °C for four 15 minutes extractions.

For each of these steps de-ionised water was renewed. The networks were dried overnight at 110 °C.

Dissolution test Strips of unmodified and glyoxalized BC samples (~ 2.5×1×0.060 mm) were

immersed in LiCl/DMAc solution for 2 months in an environmentally controlled room at a temperature

of 23 ± 1 °C and relative humidity of 50 ± 0.5 %. This solvent is reported to fully dissolve cellulose at

room temperature. The glyoxalized samples were weighed before and after immersion in LiCl/DMAc

solution using a Mettler Toledo AB265-S/FACT version classic analytical balance. Care was taken to

remove the excess of LiCl/DMAc before weighing the samples. The percentage of swelling (%S) was

determined after 48 hours using the equation

100%d

ds ×−

=WWW

S (1)

where Ws and Wd are the swollen and dry weights of the glyoxalized BC networks respectively. Five

samples were tested in total.

X-ray diffraction The degree of crystallinity and crystal morphology of unmodified and glyoxalized

BC networks were determined using an X-ray powder diffractometer (Philips X’Pert 1) having a 1.79 Å

cobalt X-ray source. Measurements were taken in the range 2θ = 10° to 40° using a step size of 0.04°.

Five samples were analysed for both unmodified and glyoxalised BC networks. In order to calculate the

percentage crystallinity (%χc) of BC, Segal’s method25 was utilized. This method uses the equation

100%002

amorphous002c ×

−=

III

χ (2)

to calculate the percentage crystallinity, where

I002 is the intensity of the 002 reflection plane and

Iamorphous is the intensity of the amorphous phase at 18°, if a copper X-ray source is used. For a cobalt

source the intensity at 21° is recorded. For the lateral crystal size (L(002)) determination, Scherrer’s

equation

L(002) =Kλ

β ⋅ cosθ (3)

was used, where β is the full width at half maximum of the 002 reflection, θ is the Bragg angle and a

value of K = 0.91 was used.

Thermogravimetric analysis The thermal degradation behavior of unmodified and glyoxalized BC

was investigated using thermogravimetric analysis (TGA Netzsch TG 209 F1). The samples

(approximately 5 mg) were heated from 25 to 600 °C at a heating rate of 5 °C min-1 under a flow of 30

mL min-1 nitrogen purge gas. Experiments were repeated three times to ensure repeatability. The onset

and degradation temperatures were obtained from the first derivative of the loss weight as a function of

temperature.

Water/air contact angle measurements Advancing and receding contact angles were measured on

unmodified and glyoxalized BC networks using a sessile drop method (DSA 10 Mk 2, Krüss) at room

temperature. Droplets of approximately 20 µL of ultra pure water were placed on the surface of

unmodified and glyoxalized BC networks using a motorized syringe and the drop volume was increased

at a rate of 6.32 µL min-1. Images of these sessile drops were processed using DSA software version

1.80.1.12. Measurements were repeated at 5 different positions on the samples.

Streaming zeta-potential measurements The zeta-potentials of unmodified and glyoxalized BC were

measured using an electrokinetic analyser (EKA, Anton Paar) based on a streaming potential method. In

order to exclude any overlaying effects, due to the swelling of unmodified and glyoxalized BC networks

or extraction of water-soluble components, the pH dependent zeta-potentials were measured after a time

dependent measurement was conducted. During the time dependent measurements, the samples were

equilibrated in 1 mM KCl electrolyte solution by means of a single long time streaming zeta-potential

measurement at 20°C. A pH-dependent zeta-potential measurement was then conducted by changing the

pH of the electrolyte solution via the titration of 0.1 N HCl or KOH, using a titration unit (RTU, Anton

Paar). The swelling behaviour was also determined from the kinetic parameters ζ0 (initial zeta potential

at first contact with water) and ζ∞ (final zeta potential, asymptotic approach). The relative change in the

zeta-potential was found using the equation

0

0 -ξξξ

ξ ∞=Δ (4)

It was found that this quotient is proportional to the water uptake at 100% relative humidity (the

sorption capacity) of natural fibres.

Water absorption capacity Unmodified and glyoxalized BC networks (~ 5×1×0.060 mm) were

conditioned for 7 days at 23 ± 1 °C and a relative humidity of 50 ± 0.5 % and immersed in de-ionised

water for 48 hours. The samples were weighed after 2, 5, 11, 24 and 48 hours using a top pan balance.

Care was taken to remove excess water before weighing the samples. Water absorption capacity

(%WAC) was determined using the equation

100%0t

0tt ×−

=www

WAC (4)

where wt0 and wt are the weights of the samples before and after water immersion respectively. Five

samples were tested for each material.

Mechanical properties Unmodified and glyoxalized BC strips (~ 20×1×0.060 mm) were secured to

20 mm gauge length testing cards using a two-part cold curing epoxy resin (Araldite®). Tensile tests

were conducted using a tensile testing machine (Instron 2511-111). The full-scale load and the

crosshead speed used were 50 N and 0.5 mm min-1 respectively. The machine compliance was

determined and found to be 4.42 × 10-3 mm N-1. All mechanical data were adjusted to take into account

the machine compliance. Dry and wet mechanical tests were conducted at 23 ± 1 °C and a relative

humidity of 50 ± 0.5 %. The samples were pre-conditioned under the same environmental conditions for

24 hours prior to testing. A total of 10 samples were tested for each material. Sample widths and

thicknesses were determined using an optical microscope and a micrometer, respectively.

For wet mechanical properties determination, unmodified and glyoxalized BC networks were

immersed in de-ionised water for 48 hours using the previously stated environmental conditions. The

samples were then removed from the water and secured on paper testing cards using cyanoacrylate glue.

The fracture surfaces of dry and wet unmodified and glyoxalized BC networks deformed in tension

were observed using a scanning electron microscope (FEG-SEM). The acceleration voltage used was 5

kV. Prior to SEM imaging, the samples were fixed onto metal stubs using carbon tabs and gold coated at

40 mA for 2 min.

Raman spectroscopy In order to follow the molecular deformation of unmodified and glyoxalized BC

networks a Raman spectrometer (Renishaw system-1000) coupled with an optical microscope and a

near infra-red laser (785 nm) was used. The laser was focussed to a 1-2 µm spot using a ×50

magnification long working distance lens. Dry unmodified and glyoxalised BC strips were mounted on

paper testing cards using a two-part cold-curing epoxy resin (Araldite®). The samples were deformed in

tension using a customized deformation rig (Deben Microtest). The compliance of the deformation rig

was determined to be 3.14 × 10-4 mm N-1. All mechanical data were adjusted to take into account the

machine compliance. The gauge length of the samples and the full scale of the load cell used were 20

mm and 2 kN respectively. During the deformation, the strain was increased incrementally by 0.025 %

and then 0.1 % at an elongation rate of 0.033 mm min-1. A Raman spectrum was recorded at each

increment using an exposure time of 30 s and 4 accumulations. The peak positions of the Raman band

initially located at approximately 1095 cm-1 were fitted using a mixed Gaussian/Lorentzian function and

an algorithm suggested by Marquardt26. Experiments were repeated 3 times for unmodified and

glyoxalized BC networks.

The molecular deformation of samples in the wet state was also investigated. Unmodified and

glyoxalized BC networks were immersed in de-ionised water for 48 hours. The samples were then

removed from the water and secured on paper testing cards using cyanoacrylate glue. During sample

preparation and molecular deformation the BC networks were regularly wetted. Two samples were

tested for each material.

Results and discussion

Glyoxal can react with the hydroxyl groups of cellulose either by an acetalyzation reaction or by

glyoxalization27. They can form either or both acetal and hemiacteal linkages during the curing

process28. In order to check if cross-linking occurred, unmodified and glyoxalized samples were

immersed in LiCl/DMAc for 2 months. Unmodified BC samples were found to be fully dissolved after

24 hours whereas glyoxalized BC samples did not dissolve, but were in a swollen state. After 48 hours,

the percentage of swelling was found to be 30 ± 4 %. Since dissolution of the glyoxalized BC networks

did not take place, it is very likely that cross-linking occurs between BC layers after this treatment. This

dissolution method has been already used to assess the extent of cross-linking in chemically modified

cotton, with cuprammonium hydroxide as a solvent29.

Figure 1 reports typical X-ray diffraction patterns for dry unmodified and glyoxalized BC networks.

It can be seen that the crystal structure is unchanged after glyoxalization since (101), (

101−

) and (002)

diffraction planes are still observed. Table 1 reports percentage crystallinity values and crystal

dimensions for dry unmodified and glyoxalized BC networks. No significant difference between dry

unmodified and glyoxalized BC networks is observed. This result suggests that cross-linking is likely to

occur only at the surface of BC fibrils, and in amorphous regions, meaning that heterogeneous

modification of BC has occurred. Hydroxyl groups located inside crystal regions are believed to be

inaccessible to chemical modification30.

Figure 2 reports the thermal degradation behavior for unmodified and glyoxalized BC networks.

The thermal behavior indicates that cross-linking has occurred, since a higher onset and peak

degradation temperature is observed. An initial weight loss was observed in the temperature range of 0

to 200 °C, which is due to water loss. No significant differences between unmodified and glyoxalized

BC networks were observed in this region, indicating that there is no difference in moisture content

between these samples (see insert). This means that if a change of mechanical or stress-transfer

properties is observed in the dry state, it should be due only to the presence of cross-linking, and not

from a plastificization effect due to the presence of moisture. A second weight loss is visible from 200

to 375 °C which is attributed to the degradation of cellulose. Table 2 reports the onset and peak

degradation temperatures for unmodified and glyoxalized BC networks. In agreement with previous

reports31, 32, there is only a slight decrease in the onset and peak degradation temperatures after

glyoxalization, meaning that this process does not dramatically decrease thermal stability.

Glyoxal has been previously found to render chitosan hydrophobic33. This suggests that it ought to be

possible to also render BC networks hydrophobic. Therefore water/air contact angle measurements were

conducted. Advancing contact angles were found to be 17.41 ± 1.74 ° and 29.59 ± 1.79 ° for unmodified

and glyoxalized BC networks respectively. A receding contact angle of 9.93 ± 0.64 ° was also measured

for glyoxalized BC networks. We could not measure the receding value for unmodified BC networks

since the material wetted before a measurement could be made, but it is likely to be low. These results

suggest a mild increase in the hydrophobicity of the BC networks, but not to the same extent as seen for

chitosan33.

Figure 3 shows the streaming zeta-potential of unmodified and glyoxalized BC as a function of pH.

The negative zeta-potential plateau and isoelectric point (iep) at acidic pH indicate that the surface of

unmodified and glyoxalized BC networks are acidic. In addition to this, the zeta-potential plateau for

glyoxalized BC networks was shifted to more negative value, indicating that its surface is more

hydrophobic than the unmodified BC networks. The hydrophobic surface of glyoxalized BC (also

confirmed by contact angle measurements) would lead to more adsorption by electrolyte ions compared

to unmodified BC, as water molecules cannot as readily adsorb due to the hydrophobic nature of the

surface of glyoxalized BC networks. A shift in the iep of glyoxilized BC to lower pH was also observed.

This might be due to the fact that the glyoxilization reaction introduced more free acidic functional

groups to the surface. More dissociable acidic functional groups implies lower iep. We also calculated

Δζ from ζ=f(t) for unmodified and glyoxalized BC networks (see figure 3b). Values of 0.255 and 0.184

were found for unmodified BC and glyoxalized BC, respectively. The higher value of Δζ indicates that

unmodified BC networks have a higher water absorption capacity compared to glyoxalized BC

networks. This is not surprising due to the hydrophilic nature of unmodified BC.

Figure 4 reports water absorption as a function of time for unmodified BC and glyoxalized BC. After

two hours unmodified and glyoxalized BC networks were saturated with water; no significant changes

were observed after 5, 12, 24 and 48 hours. A significant difference between the relative water

absorption capacities of unmodified and glyoxalized BC networks is however noted; unmodified BC is

much higher than glyoxalized BC. This result is consistent with the calculation of the variation of zeta-

potential (the lower that variation, the lower the water absorption capability) and with the contact angle

and zeta-potential measurements since a more hydrophobic surface will absorb less water than a more

hydrophilic one. Hence the presence of cross-linking must also reduce the swelling of BC networks with

water. This reduced water absorption capacity has also been observed for biodegradable plastics made

from soy bean products cross-linked with glyoxal33. This difference in water absorption capacity will

have to be taken into account when interpreting the stress-transfer and mechanical properties in the wet

state. This result is also interesting because moisture resistance is often an issue for cellulose-based

biocomposites. The use of glyoxalised BC networks could therefore help to produce biocomposites with

enhanced moisture/swelling resistance.

Figures 5a and 5b show scanning electron microscopy images of the fracture surfaces of dry

unmodified and glyoxalized BC networks deformed in tension. Clear delamination of BC layers is

observed for the unmodified networks. BC fibrils can also be seen bridging these delaminated layers

which must play a role in the fracture process. Such a delamination process does not occur for dry

glyoxalised BC networks, which must be due to the presence of chemical cross-linking between layers.

Figure 5c reports scanning electron microscope images for unmodified BC samples fractured whilst wet

(the samples are dry within the vacuum chamber of the SEM). The cross-sections of the samples appear

to be considerably reduced compared to the dry samples, and we did not observe any delamination of

the layers. Some kind of plastification effect (see arrows) is however observed. Figure 5d reports a

scanning electron microscopy image for the tensile fracture surface of a wet glyoxalized BC network.

These samples were found to maintain their dimensions, not reducing in thickness. Delamination of BC

layers is however observed for this sample.

Figure 6 reports typical stress-strain curves for dry and wet unmodified and glyoxalized BC

networks. The mechanical properties of dry unmodified and glyoxalized BC networks are also reported

in Table 3. In the dry state, one can clearly see that glyoxalization of the samples reduces stress and

strain at failure, and the work of fracture, indicating an embrittlement of the material. Indeed this

embrittlement is matched by the clean fracture surfaces observed for glyoxalized BC networks (see Fig.

5b). This embrittlement may occur due the presence of cross-linking between BC layers, which will

reduce layer-to-layer mobility, and consequently delamination. Cross-linking must also occur in the

plane of the specimen, which could reduce orientation effects previously reported for BC networks34.

No difference between Young’s modulus of unmodified and glyoxalized BC networks was however

observed, suggesting little difference in the cross-linking in the longitudinal direction. The transverse

mechanical properties could be significantly different, due to cross-linking between layers in the BC

network, but that remains a topic for future work. It is also possible that there could be a decrease of the

molecular weight of the BC networks upon glyoxalization. However we would require more evidence to

support this hypothesis.

When the BC networks are in the wet state, a large decrease of Young’s modulus and stress at failure

occurs. An increase in the strain at failure for wet unmodified BC networks compared to dry unmodified

BC networks also occurs. Young’s modulus, stress at failure and strain at failure change by 438, 1570

and 261 % respectively, when samples are in the wet state. These large significant changes could be

explained by the disruption of hydrogen bonding by competitive binding of water molecules, allowing

slippage between BC fibrils. The modulus of wet glyoxalized BC networks is reduced by 77 %,

compared to the dry state. The stress at failure does not change significantly, but the strain at failure is

increased by 137 %. The presence of crosslinks which prevent cellulose polymer chains from slipping

from one another is therefore thought to maintain the strength of the networks.

Figure 7a reports typical Raman spectra for unmodified and glyoxalized BC networks. We did not

observe any difference between spectra which suggests that the cellulose backbone has not been

affected by chemical modification, which also confirms that heterogeneous modification has occurred.

The Raman band initially located at 1095 cm-1, highlighted in Figure 7a, has been extensively used for

stress-transfer quantification in cellulose materials1, 17, 15, 34. This band is thought to be representative of

C-O ring stretching modes35 and/or the C-O-C glycosidic bond stretching36, 37.

Figures 7b and 7c report typical shifts in the position of the Raman band initially located at 1095 cm-

1 for dry unmodified and glyoxalized BC networks respectively. It was noted that the shift in this band is

higher in magnitude for dry networks compared to wet samples, in agreement with previous work on

wet cellulose nanowhisker-based composite materials38. Figure 8a reports the detailed positions of the

Raman band initially located at 1095 cm-1 as a function of strain for dry unmodified and glyoxalized BC

networks. The stress-transfer efficiency of the networks can be obtained from the gradient of a linear fit

to these data. A higher stress-transfer efficiency is observed for glyoxalised BC networks compared to

unmodified samples, which is an indication of the cross-linking that has taken place. The gradient of

these linear fits for dry unmodified and glyoxalized BC networks were -0.85 ± 0.13 cm-1 %-1 and -1.66

± 0.05 cm-1 %-1 respectively. As previously noted, less delamination occurred for glyoxalized BC

networks, which is thought to contribute to the higher stress-transfer efficiency. As a consequence of

this, less energy is likely to be dissipated throughout the structure. This is confirmed by the higher work

of fracture for dry unmodified BC networks compared to glyoxalized BC networks (Table 3).

Figure 8b reports the wet stress transfer properties of the networks, as revealed by the shift in the

position of the Raman band located at 1095 cm-1. It is clear that the presence of water within unmodified

BC networks almost completely suppresses their stress-transfer efficiency. This effect has also been

observed for cellulose whisker-based nanocomposites exposed to water34. This suggests that water has

fully penetrated the BC networks, and therefore acts as a plasticizer, disrupting hydrogen bonding

between fibrils, and consequently reducing local molecular deformation and therefore stress-transfer

within the network. The stress-transfer efficiency of wet glyoxalized BC networks is however preserved

by cross-linking; the linear fit to these data was found to be 1.32 cm-1%-1, which is only slightly lower

than for dry glyoxalized BC networks (1.66 cm-1 %-1). This explains why the mechanical properties of

glyoxalized BC networks, particularly strength, were hardly affected by the presence of water (see Table

3). Differences in the stress-transfer efficiencies of dry and wet glyoxalized BC networks must be due to

the removal of hydrogen bonding after immersion in water. Water absorption experiments showed that

glyoxalized BC networks absorb much less water than unmodified BC networks, and so this might also

explain why stress-transfer is preserved for wet glyoxalized BC networks.

Conclusions

This study has shown that it is possible to crosslink BC networks via glyoxalization. The

glyoxalization process was found to not significantly change the degree of crystallinity and crystal

morphology of the BC, meaning that heterogeneous modification of cellulose occurred.

Thermogravimetric analysis showed that the onset and peak degradation temperatures were significantly

reduced after glyoxalization. Contact angle and zeta-potential measurements showed that glyoxal

rendered the BC networks hydrophobic. In the dry state, Young’s modulus was found not to be

significantly changed after glyoxalization. Stress and strain at failure, and work of fracture, were found

to however reduce upon modification. The reduced work of fracture values were supported by tensile

fracture surface observations using scanning electron microscopy, suggesing that delamination of the

layered structure of glyoxalized BC networks also reduced, owing to the presence of cross-linking. In

the wet state, the mechanical properties of unmodified BC networks were reduced due to the disruption

of hydrogen bonding. The mechanical properties, particularly the strength, of wet glyoxalized BC

networks did not however reduce, compared to the dry state, again due to the presence of cross-linking.

This was also supported by scanning electron microscopy imaging. Consequently glyoxal cross-linking

offers dimensional and mechanical stability to the BC networks when exposed to moisture. Water

absorption capacity experiments clearly showed that glyoxalized BC networks do not absorb as much

water as unmodified BC networks. Zeta–potential measurements also confirmed this result. These

findings could have implications for the design of biocomposites generated from BC with increased

moisture resistance. Finally, Raman spectroscopy revealed the stress-transfer mechanisms of

unmodified and glyoxalized BC networks both in the dry and wet states. Micromechanical

characterization showed that glyoxalization of BC networks results in a higher stress-transfer efficiency.

This was found to be also true in the wet state. This greater stress-transfer efficiency of glyoxalized BC

networks is thought to be due to the presence of cross-linking, but also to difference in the relative water

absorptions of the networks. The use of glyoxal therefore provides a facile approach to modify the

stress-transfer properties of BC networks, which could be used for a range of applications.

Acknowledgements

The authors would like to thank the EPSRC for funding the PhD studentship under Grant GR/F028946

and EP/F032005/1. Thanks are extended to Dr. Henning Althoefer from BASF for useful advises.

References

(1. Ranby, B. G. Arkiv for Kemi 1952, 4, (3), 241-&. 2. Sturcova, A.; Davies, G. R.; Eichhorn, S. J. Biomacromolecules 2005, 6, (2), 1055. 3. Brown, A. J. Journal of Chemical Society 1986, (49), 172. 4. Brown, A. J. Journal of Chemical Society 1986, (49), 432. 5. Jonas, R.; Farah, L. F. Biodegradable Polymers and Macromolecules 1998, 59, (1-3), 101. 6. Keshk, S. M. A. S.; Razek, T. M. A.; Sameshima, K. African Journal of Biotechnology 2006, 5, (17), 1519. 7. Nishi, Y.; Uryu, M.; Yamanaka, S.; Watanabe, K.; Kitamura, N.; Iguchi, M.; Mitsuhashi, S. Journal of Materials Science 1990, 25, (6), 2997. 8. Kim, Y.; Jung, R.; Kim, H. S.; Jin, H. J. Current Applied Physics 2009, 9, (1 SUPPL.), S69. 9. Gindl, W.; Keckes, J. Composites Science and Technology 2004, 64, (15 SPEC. ISS.), 2407. 10. Huang, Y.; Wan, Y.; Hu, L.; He, F.; Wang, Y. Fuhe Cailiao Xuebao/Acta Materiae Compositae Sinica 2008, 25, (6), 140. 11. Yano, H.; Sugiyama, J.; Nakagaito, A. N.; Nogi, M.; Matsuura, T.; Hikita, M.; Handa, K. Advanced Materials 2005, 17, (2), 153-155. 12. Gea, S.; Torres, F. G.; Troncoso, O. P.; Reynolds, C. T.; Vilasecca, F.; Iguchi, M.; Peijs, T. International Polymer Processing 2007, 22, (5), 497. 13. Lee, K.-Y.; Blaker, J. J.; Bismarck, A. Composites Science and Technology. 2009.

14. Sakurada, I.; Nukushina, Y.; Ito, T. Journal of Polymer Science 1962, 57, (165), 651-660. 15. Nishino, T.; Takano, K.; Nakamae, K. Journal of Polymer Science Part B: Polymer Physics 1995, 33, (11), 1647-1651. 16. Hsieh, Y. C.; Yano, H.; Nogi, M.; Eichhorn, S. J. Cellulose 2008, 15, (4), 507. 17. Guhados, G.; Wan, W.; Hutter, J. L. Langmuir 2005, 21, (14), 6642. 18. Quero, F.; Nogi, M.; Yano, H.; Abdulsalami, K.; Holmes, S. M.; Sakakini, B. H.; Eichhorn, S. J. 2009, 2, (1), 321. 19. Nogi, M.; Yano, H. Advanced Materials 2008, 20, (10), 1849. 20. Klemm, D., Philipp B., Heinze T., Heinze U., Wagenknecht W. 1998, 2. 21. Ramires, E. C.; Megiatto Jr, J. D.; Gardrat, C.; Castellan, A.; Frollini, E. 101, (6), 1998. 22. Miyata, T.; Kurokawa, K.; Van Ypersele De Strihou, C. Journal of the American Society of Nephrology 2000, 11, (9), 1744. 23. Welch, C. M.; Forthright Danna, G. Textile Research Journal 1982, 52, (2), 149. 24. Hestrin, S.; Schramm, M. Biochem. J. 1954, 58, 345-352. 25. Segal, L., Creely, J.J., Martin, A.E., Jr, Conrad, C.M. Textile Research Journal 1959, 29, 786-794. 26. Marquardt, D. W. Society for Industrial and Applied Mathematics 1963, 11, (2), 431. 27. Head, F. S. H. Journal of the Textile Institute Transactions 1958, 49, (7), 345 - 356. 28. Schramm, C.; Rinderer, B. 2000, 72, (23), 5829. 29. Reeves, W. A.; Drake, G. L. J. R.; McMillan, O. J. J. R.; Guthrie, J. D. Textile Research Journal 1955, 25, (1), 41. 30. Krassig. 1993. 31. Lee, E.; Kim, S. Fibers and Polymers 2004, 5, (3), 230. 32. Hyung-Min, C.; Jung Hyun, K.; Sangmoo, S. Journal of Applied Polymer Science 1999, 73, (13), 2691-2699. 33. Gupta, K. C.; Jabrail, F. H. 2006, 341, (6), 744. 34. McKenna, B. A.; Mikkelsen, D.; Wehr, J. B.; Gidley, M. J.; Menzies, N. W. Cellulose 2009, 16, (6), 1047-1055. 35. Wiley, J. H.; Atalla, R. H. Carbohydrate Research 1987, 160, 113-129. 36. Edwards, H. G. M.; Farwell, D. W.; Webster, D. Spectrochimica Acta Part a-Molecular and Biomolecular Spectroscopy 1997, 53, (13), 2383-2392. 37. Gierlinger, N.; Schwanninger, M.; Reinecke, A.; Burgert, I. Biomacromolecules 2006, 7, (7), 2077-2081. 38. Rusli, R.; Shanmuganathan, K.; Rowan, S. J.; Weder, C.; Eichhorn, S. J. Biomacromolecules 2010, 11, (3), 762-768.

LIST OF TABLES

Table 1 Percentage crystallinity (χχ c) and lateral crystal sizes (L) for dried unmodified and

glyoxalized BC networks

Table 2 Onset and peak degradation temperatures for unmodified and glyoxalized BCnetworks.

Table 3 Mechanical properties for dry and wet unmodified and glyoxalized BC networks;E -

Young’s, modulus, σf - stress at failure, εf - strain at failure and G - work-of-fracture.

LIST OF FIGURES

Figure 1 Typical X-ray diffraction patterns for dry unmodified and glyoxalized BC networks.

Figure 2 Typical thermal degradation profiles for unmodified and glyoxalized BC networks. Error

bars are standard deviations from the mean. The insert represents the thermal behavior

from 25 ºC to 200 ºC.

Figure 3 (a) Streaming zeta-potentials of unmodified and glyoxilized BC as a function of pH.; (b)

Streaming zeta-potentials of unmodified and glyoxilized BC as a function of time.

Figure 4 Relative water absorption as a function of time for unmodified and glyoxalized BC.

Figure 5 Scanning electron microscope images of fracture surfaces for dry (a) unmodified and (b)

glyoxalized BC networks and wet (c) unmodified and (d) glyoxalized BC networks.

Figure 6 Typical stress-strain curves for dry and wet unmodified and glyoxalized BC networks.

Figure 7 (a) Typical Raman spectra for dry unmodified and glyoxalized BC networks; Typical

shifts in the position of the Raman band initially located at 1095 cm-1 towards a lower

wavenumber for a dry (b) unmodified BC network and (c) a glyoxalized BC network.

Figure 8 Detailed typical shifts in the position of the Raman band located at 1095 cm-1 for (a) dry

unmodified and glyoxalized BC networks and (b) wet unmodified and glyoxalized BC

networks. Errors are standard deviations from the mean for 2 samples (BC1 and BC2)

Material χχ c (%) L (002) (Å) d (002) (Å) Unmodified BC 94.31 ± 1.21 62.97 ± 2.12 7.82 ± 0.05 Glyoxalized BC 93.85 ± 0.18 63.56 ± 0.47 7.85 ± 0.04

Table 1

Material Onset degradation temperature (ºC)

Peak degradation temperature (ºC)

Unmodified BC 319.00 ± 1.12 334.58 ± 3.82 Glyoxalized BC 299.30 ± 0.78 329.58 ± 2.89

Table 2

Material E (GPa) σf (MPa) εf (%) G (MJ m-3) Dry unmodified BC 10.12 ± 1.53 165.14 ± 33.86 2.57 ± 0.63 2.52 ± 0.90 Dry glyoxalized BC 10.67 ± 1.36 71.04 ± 36.04 0.64 ± 0.32 0.26 ± 0.24 Wet unmodified BC 1.88 ± 0.24 9.89 ± 1.47 9.28 ± 1.86 5.18 ± 0.62 Wet glyoxalized BC 6.04 ± 1.41 76.82 ± 21.23 1.52 ± 0.66 1.29 ± 0.84

Table 3

10 15 20 25 30 35 400

10000

20000

30000

40000

50000

60000

70000

80000

Unmodified BC Glyoxalized BC

Inte

nsity

(Arb

itrar

y un

its)

Diffraction angle 2θ (°)

(002)

(101)

(101)

Figure 1

0 100 200 300 400 500 6000

20

40

60

80

100

50 100 150 20096

97

98

99

100

Wei

ght l

oss

(%)

Temperature (oC)

Unmodified BC Glyoxalized BC

Figure 2

(a)

2 4 6 8 10 12-12

-10

-8

-6

-4

-2

0

2

4

Unmodified BC Glyoxalized BC

Stre

amin

g ζ-

pote

ntia

l (m

V)

pH

(b)

0 100 200 300 400 500 600 700 800

-12

-10

-8

-6

-4

-2

0

2

4

Unmodified BC Glyoxalized BC

Stre

amin

g ζ−

pote

ntia

l (m

V)

Time (Minutes)

Figure 3

0 10 20 30 40 50

0

10

20

30

40

50

60

70

80

90

100

Unmodified BC Glyoxalized BC

Rel

ativ

e w

ater

abs

orpt

ion

(%)

Time (hours)

Figure 4

(a)

(b)

(c)

(d)

Figure 5

0 1 2 3 4 5 60

20

40

60

80

100

120

140

160

180

Stre

ss (G

Pa)

Strain (%)

Dry unmodified BC Dry glyoxalized BC Wet unmodified BC Wet glyoxalized BC

Figure 6

(a)

400 600 800 1000 1200 1400 16000

10000

20000

30000

40000

50000

60000

70000

80000

90000

100000

110000

120000

Inte

nsity

(Arb

itrar

y un

its)

Raman shift (cm-1)

Unmodified BC Glyoxalized BC

1095 cm-1

(b)

1040 1060 1080 1100 1120 1140 1160

10000

20000

30000

40000

50000

60000

70000

80000

90000

100000

110000

120000

Inte

nsity

(Arb

itrar

y un

its)

Raman shift (cm-1)

0% 0.5%

(c)

1040 1060 1080 1100 1120 1140 1160290000

300000

310000

320000

330000

340000

350000

360000

370000

380000

390000

Inte

nsity

(Arb

itrar

y un

its)

Raman shift (cm-1)

0% 0.5%

Figure 7

(a)

0.0 0.5 1.0 1.5 2.0 2.5 3.0-2.5

-2.0

-1.5

-1.0

-0.5

0.0

Gradient of fit =-0.85 ± 0.13 cm-1 %-1

R2=0.98

Gradient of fit =-1.66 ± 0.05 cm-1 %-1

R2=0.98

Dry unmodified BC Dry glyoxalized BC

Ram

an b

and

shift

(cm

-1)

Strain (%)

(b)

0.0 0.5 1.0 1.5 2.0 2.5 3.0-2.5

-2.0

-1.5

-1.0

-0.5

0.0

Wet unmodified BC 1 Wet unmodified BC 2 Wet glyoxolized BC 1 Wet glyoxolized BC 2

Ram

an b

and

shift

(cm

-1)

Strain (%)

Gradient of fit=-1.32 cm-1 %-1

R2=0.98

Figure 8